BACKGROUND OF THE INVENTION
1. Field of the Invention
[0001] The present invention relates to a high-performance collimator and compact spectrophotometer
capable of performing spectrophotometry without using a prism or diffraction grating.
2. Description of the Related Art
[0002] Conventionally, a method for measuring spectral intensity by guiding light in a different
direction for each wavelength using a prism or a diffraction grating, irradiating
the light to a linear sensor or the like, and measuring the output from the elements
of the linear sensor was generally used to perform spectrophotometry. However, a certain
amount of space is needed to separate and guide the diffracted light in different
directions according to wavelength when a prism, diffraction grating, or the like
is used. An unacceptably large spectrophotometer thus resulted. Other drawbacks included
the fact that the light intensity decreased due to the measured light being passed
through a slit when guided to the prism or diffraction grating, and that high-speed
measurement was difficult to accomplish because the storage time required for the
linear sensor was considerable.
[0003] Several methods featuring a linear variable filter (occasionally referred to hereafter
as LVF) have been suggested and developed with the aim of overcoming the abovementioned
drawbacks. For example, the technique disclosed in Japanese Patent Application Laid-open
No. H5-322653, and the technique disclosed in US Patent No. 5,872,655 are widely known,
and a spectrophotometer that uses a different system of linear variable filter is
disclosed in US Patent No. 6,057,925 and is commercially available. In this techniques,
diffracted light emitted from a linear variable filter is focused on a linear sensor
by inserting an optical system for an erect same-size image between the linear variable
filter and linear sensor, and a GRIN (Gradient Index) lens or a Micro Lens Array is
used as a compact focusing system for the erect same-size image.
[0004] The following problems nonetheless exist in the system disclosed in Japanese Patent
Application Laid-open No. H5-322653, and in the method disclosed in US Patent No.
5,872,655. Specifically, multiple reflections occur between the linear variable filter
and linear sensor due to these methods being configured such that the linear variable
filter is attached to the linear sensor, and the spectral characteristics thus deteriorate.
[0005] While these problems are eliminated in the method disclosed in US Patent No. 6,057,925,
other problems nonetheless occur. Specifically, the GRIN lens comprises a total of
28 cylindrical lenses disposed in two rows. Consequently, when a surface image similar
to a linear variable filter is projected, exactly 28 peaked irregularities occur in
the output of the linear sensor because a composite image made by 28 cylindrical lenses
is focused on the linear sensor. The dimensional accuracy of the spectral wavelength
output will thereby decrease even when the positional accuracy of the spectral wavelength
is enhanced.
[0006] Meanwhile, mechanical collimators are conventionally used for transmitting light
in a parallel fashion. Specifically, a large-scale collimator is used in large-scale
optical systems such as edge locators and width meters for web-shaped measurement
objects. This kind of mechanical collimator may also be used in the transmission of
light between a linear variable filter and a linear sensor. However, such collimators
have bulky structures, and it was believed to be impossible to manufacture a compact
product that satisfies high resolution requirements and is capable of being used in
transmission of light between a linear variable filter and a linear sensor.
[0007] Aiming to develop a method for overcoming the aforementioned problems, the inventors
have succeeded in obtaining spectral results having high wavelength resolution and
devoid of any irregularities by placing a fiber optic sheet (hereafter abbreviated
as FOP) between the linear variable filter and the linear sensor, or at the forward
surface of the linear variable filter, and have filed for a patent as Patent Application
2001-078176 (hereafter referred to as "the prior application"). An FOP having high
directivity (NA = 0.35) is used especially for enhancing wavelength resolution in
the embodiments of this invention.
[0008] A structural diagram thereof is depicted in Fig. 16. Fiber optic sheets 33a and 33b
are provided above the sensor package 32 of a linear sensor 31 in the manner shown
in the figure, and a linear variable filter 34 is bonded to the fiber optic sheet
33a on the top thereof. The space between the linear sensor 31 and the fiber optic
sheet 33b is filled with a transparent resin 35, and is approximately 0.01 mm wide.
The numerical aperture (NA) of the fiber optic sheets 33a and 33b is assumed to be
1 in this embodiment.
[0009] The reason that the fiber optic sheets 33a and 33b are divided into two layers is
that the fiber optic sheet 33a is used instead of the surface cover glass of the sensor
package 32 of the linear sensor 31; and when this is unnecessary, a single fiber optic
sheet may be used, or 33a and 33b may be formed in an integrated manner.
[0010] For light entering the linear variable filter 34 from the upper portion of the figure,
only light with a wavelength determined by the entry location thereof in the linear
variable filter 34 is transmitted, spectrally divided according to the location of
the linear variable filter 34, guided by the fiber optic sheets 33a and 33b, and directed
through the transparent resin 35 to the corresponding pixel of the linear sensor 31.
Spectral measurement can thus be performed by processing the output of each pixel
of the linear sensor 1.
[0011] Because the linear variable filter 34 and fiber optic sheet 33a are attached, there
is no light diffusion in the space therebetween, but a slight degree of light diffusion
still occurs between the fiber optic sheet 33b and the linear sensor 31 even if the
numerical aperture of the fiber optic sheets 33a and 33b is equal to 1. However, this
does not present much of a problem, since the interval thereof is approximately 0.01
mm. The light transmission rate from the linear variable filter 4 to the linear sensor
1 is approximately 60-70% in this embodiment, which is apparently by no means inferior
when compared with the conventional example, in which a linear variable filter and
linear sensor are adjacent.
[0012] Problems nonetheless exist in this method as well. Specifically, the effective waveband
of the FOP in which the actual NA is 0.35, ranges only from 400 to 800 nm. Consequently,
high-resolution spectral results cannot be obtained for longer wavelengths (800 nm
or greater). At the same time, it is said that linear variable filters used for wave
length from 0.4 µm to 20 µm can be manufactured .
[0013] To use an FOP at greater wavelengths, a fiber for the waveband thereof must be newly
manufactured. It is costly and impractical to make an optical fiber corresponding
to each wavelength. Another drawback is that light transmissivity is attenuated when
NA = 0.35, even in a visible-range FOP.
SUMMARY OF THE INVENTION
[0014] An object of the present invention, which was developed in view of the above situation,
is to provide a wavelength-independent compact collimator having high wavelength resolution,
and to provide a compact spectrophotometer that uses this collimator and is capable
of spectral measurement with high speed and high accuracy.
[0015] The first invention for achieving the aforementioned object is a collimator in which
the light transmission path is air.
[0016] Because the light transmission path consists of air in the present invention, it
is possible to obtain a wavelength-independent collimator having little attenuation
of infrared light. Specifically, the optical path length for a usual collimator is
approximately 0.5-5 mm, and there is almost no need to consider wavelength dependence
when the device is used from the ultraviolet to far infrared regions.
[0017] The second invention for achieving the aforementioned object is a collimator formed
by alternately stacking first metal sheets having holes and second metal sheets devoid
of holes, pressing opposite sides thereof with pressing sheets, integrating these
parts by diffusion bonding based on thermocompression, and cutting the portions corresponding
to the portions where the holes exist in the first metal sheets, in the stacking direction
of the metal sheets.
[0018] The collimator formed according to the present invention is configured such that
a plurality of holes whose width is equal to the thickness of the aforementioned first
metal sheets are formed in parallel, leaving open an interval equal to the thickness
of the aforementioned second metal sheets. Specifically, a collimator is formed having
a number of openings equal to the number of first metal sheets. The first metal sheets
and second metal sheets are bonded by diffusion bonding based on thermocompression
bonding, and can therefore be made sufficiently thin (specifically, their thickness
may be on the order of tens of micrometers), and a collimator can thus be formed in
which a large number of holes tens of micrometers wide are disposed at intervals of
tens of micrometers.
[0019] The third invention for achieving the aforementioned object is a two-dimensional
collimator formed by alternately stacking first metal sheets having parallel holes
in a plurality of rows and second metal sheets devoid of holes, pressing opposite
sides thereof with pressing sheets, integrating these parts by diffusion bonding based
on thermocompression bonding, and cutting the portions corresponding to the portions
where the holes exist in the first metal sheets, in the stacking direction of the
metal sheets.
[0020] In the present invention, the aforementioned parallel holes are arranged at an interval
in one direction, and the holes are arranged according to the number of first metal
sheets in the direction orthogonal thereto, as described in the aforementioned first
means. It is thus possible to construct a two-dimensional collimator having the same
structure as the aforementioned second invention.
[0021] The fourth invention for achieving the aforementioned object is a collimator in which
metal sheets with holes whose length is sufficient to cover the width direction of
the holes in the aforementioned first metal sheets in the cut portions are used instead
of the second metal sheets in the aforementioned second and third inventions, and
is otherwise manufactured by the method described in claim 2 or 3.
[0022] Metal sheets having holes with a length sufficient to cover the width direction of
the holes in the aforementioned first metal sheets in the cut portions are used instead
of the second metal sheets in the present invention. Specifically, holes are opened
in the cut portions of the second metal sheets, the length of the holes in the cutting
planes is equal to or greater than the length of the holes in the cutting planes of
the first metal sheets, and the holes in the second metal sheets cover the holes in
the first metal sheets in the cutting planes.
[0023] Because of this, cutting may be performed until the cut portion reaches the holes
in the first metal sheets during cutting, and there is no need to cut the portions
of second metal sheets that comprise the barriers of the collimator holes (the holes
are formed in advance) in the second metal sheets. These portions thus experience
no deformation from the cutting force or heat during cutting.
[0024] The fifth invention for achieving the aforementioned object comprises any of the
aforementioned inventions 2 through 4, wherein the holes formed in the first metal
sheets have grid members formed in the lengthwise direction thereof.
[0025] The holes formed in the first metal sheets in the present means have grid members
formed in the lengthwise direction, specifically, in the direction orthogonal with
respect to the cutting planes thereof. These grid members serve as braces for supporting
the second metal sheets that comprise the barriers of the collimator holes, and prevent
deformation of the second metal sheets when the collimator is completed, as later
described in detail in the embodiments section using the figures. It is thus possible
to obtain a collimator whose holes have a regular shape.
[0026] The sixth invention for achieving the aforementioned object is a collimator comprising
a capillary plate.
[0027] The capillary plate is conventionally made as an image intensifier, but is used as
a collimator in the present means. The capillary plate is manufactured as follows.
Dual-tube glass having different components for the inner tube and outer tube thereof
is extended to an appropriate girth, disposed in a maximally packed configuration,
and heat-treated to fusion-bond the glass tube. The inner tube of the dual tube is
melted with acid after cross-sectional cutting, and a cavity is formed. Blackening
is performed for the capillary plate formed in this manner, so that the surface and
cavity sidewalls thereof are endowed with zero reflectivity.
[0028] In this arrangement, as light entering from one side of the capillary plate passes
through the cavity, light that is parallel to the axis of the cavity proceeds straight
through, and angled light is absorbed and attenuated every time it reflects off the
cavity walls, failing to reach the surface on the opposite side when the cavity length
is considerable. The capillary plate configured in this manner can thus be used as
a collimator.
[0029] The seventh invention for achieving the aforementioned object comprises a spectrophotometer
having a linear variable filter, a linear sensor, and a collimator for transmitting
spectrally divided light emitted from the linear variable filter to the linear sensor,
disposed between the linear variable filter and linear sensor, wherein the collimator
used is any of the collimators according to any of the aforementioned first through
sixth means.
[0030] The configuration of the present invention is generally identical to the configuration
of the invention of "the prior application" described in the prior art, but differs
in that a collimator according to any of the aforementioned inventions 1 through 6
is used instead of a fiber optic plate (FOP). An FOP is also a type of collimator,
but the light propagates inside a fiber, and the wavelength dependence of the refractive
index of the fiber affects the propagation characteristics. In contrast, the light
transmission path in any of the collimators used in the present invention consists
of air, so there is no wavelength dependence, and light ranging from infrared light
to ultraviolet light can be spectrally divided.
[0031] An objective element linear sensor, which is an example of a spectrometer in which
a linear variable filter is used, is extremely small, having a length of 12.5 mm and
comprising 256 wavelength detecting elements, each 50 µm wide (width: 2500 µm). In
addition, the pitch dimension of the collimator cavity should preferably be about
10-100 µm because the linear variable filter and the linear sensor have the same dimensions
(1:1).
BRIEF DESCRIPTION OF THE DRAWINGS
[0032]
Fig. 1 is a diagram depicting a conceptual outline of the collimator according to
the first example of the embodiments of the present invention;
Fig. 2 is a diagram for describing an example of the manufacturing method for the
collimator depicted in Fig. 1;
Fig. 3 is a diagram for describing an example of the manufacturing method for the
collimator depicted in Fig. 1;
Fig. 4 is a diagram depicting an example of the structure of members for manufacturing
a two-dimensional collimator, and an example of the cross-section of the two-dimensional
collimator;
Fig. 5 is a diagram depicting an example of the structure of members for manufacturing
a grid as a reinforcing member, and an example of the cross-section of a one-dimensional
collimator having the grid;
Fig. 6 is a diagram depicting the structure of members for preventing deformation
during cutting;
Fig. 7 is a diagram depicting a structural outline of a spectrophotometer as an example
of an embodiment of the present invention;
Fig. 8 is a diagram depicting an outline of a capillary plate used as a collimator;
Fig. 9 is a diagram depicting an outline of a device used for investigating the characteristics
of the spectrophotometer according to an embodiment of the present invention;
Fig. 10 is a diagram depicting the results of direct spectral decomposition of light
emitted from a parallel light source;
Fig. 11 is a diagram depicting the results of spectral decomposition of emitted light
from a parallel light source through a didynium filter;
Fig. 12 is a diagram depicting the results of spectral decomposition of emitted light
from a parallel light source via a diffusion plate;
Fig. 13 is a diagram depicting the results of spectral decomposition of emitted light
from a parallel light source via a diffusion plate and didynium filter;
Fig. 14 is a diagram depicting the results of spectral decomposition of light transmitted
through a didynium filter using an infrared light source as a parallel light source;
Fig. 15 is a diagram depicting the results of spectral decomposition of emitted light
from a parallel light source through a didynium filter in the invention of "the prior
application", in which an FOP is used as a collimator; and
Fig. 16 is a diagram depicting an example of the configuration of the spectral device
of the invention of "the prior application".
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0033] Embodiments of the present invention will be described hereafter using the figures.
Fig 1 is a diagram depicting a conceptual outline of the collimator according to the
first example of the embodiments of the present invention. In Fig. 1, (a) is a plan
view; (b) is a front view; (c) is an A-A cross-sectional view; and (d) is a B-B cross-sectional
view. Because this figure is a conceptual diagram for describing the structure, the
dimensions shown do not correspond to actual dimensions.
[0034] As is apparent from the diagram, the collimator is obtained by the alternate stacking
of metal sheets 1 (40 µm thick) having holes 4 with a width of 2200 µm in the center
thereof, and metal sheets 2 (10 µm thick) without holes (here, "the metal sheets 1
with the holes 4" describes in the state in which they exist before being cut in the
manner described below, the upper portion and the bottom portion of the metal sheets
1 in the figure (in the finished product) are not connected with each other). Both
sides are held by pressing sheets 3 that are 2 mm thick. These metal sheets and pressing
sheets are bonded by means of diffusion bonding based on thermocompression bonding.
[0035] The portions with the vertical through-holes 4 (40 µm× 2000 µm) thus become light-transmitting
portions, the metal sheets 2 serve as barriers between adjacent holes 4, and the light
collimated to a width of 40 µm ultimately passes through. The metal thin sheet used
may be any photoetchable metal thin sheet as long as this film can be stacked and
diffusion-bonded using thermocompression bonding. In this case, however, an SUS sheet
is used. The sheet has high strength and is comparatively inexpensive and widely available.
Aluminum is another effective material, but SUS is superior in strength. The structure
of the portion depicted by the dotted line in the figure is identical to the portions
on the right and left thereof, and is thus not shown in the diagram. 256 metal sheets
1 and 255 metal sheets 2 are stacked, and 256 light passages are formed in this embodiment.
[0036] Because of the novelty of this collimator, an example of the manufacturing method
thereof will be described. As depicted in Fig. 2, SUS thin sheets 1 with a length
of 100 mm, a width of 8 mm, and a thickness of 40 µm; SUS thin sheets 2 with a length
of 100 mm, a width of 8 mm, and with a thickness of 10 µm; and an SUS sheet 3 with
a length of 100 mm, a width of 8 mm, and a thickness of 2 mm are prepared, and holes
4 measuring 40 µm×2200 µm are formed in the central portions of the SUS thin sheet
1 by photolithography and etching. Photolithography and etching are also used for
each of the SUS thin sheets 1 and the SUS thin sheets 2, electric discharge machining
is used for the SUS sheets 3 to make two holes 5 having a diameter of 2 mm in them.
The reason that etching is used as a processing method is to eliminate the occurrence
of weld flashing.
[0037] A 40-µm SUS thin sheet 1 is next placed over the 2-mm-thick SUS sheet 3, and a 10-µm
SUS thin sheet 2 is stacked thereon. 40-µm and 10-µm SUS thin sheets are then alternately
stacked. In this example, 256 40-µm SUS sheets 1 and 255 10-µm SUS sheets 2 are stacked,
and a 2-mm-thick SUS sheet 3 is placed thereon. The sheets are then aligned using
the holes 5 with the 2-mm diameters.
[0038] The stacked sheets must be joined to one another, since they are not secured in this
state. A thermocompression bonding technique may be used herein to join together the
contacting surfaces of the SUS sheets. To accomplish this, pressure is applied to
the stacked portion with the aid of pressing sheets (using material not joined with
the SUS) from above and below the stacked portions, the assembly in this state is
placed in a vacuum heating furnace, the temperature is raised from room temperature
to 1000°C and held at this level, and the temperature is reduced when diffusion bonding
is considered to be completed. This step takes about 24 hours. In this manner, a bonded
multilayered sheet as depicted in Fig. 3 is completed. In Fig. 3, (a) is a plan view,
and (b) is a side view.
[0039] The bonded multilayered sheet is then cut in the stacking direction thereof. The
cutting location for cutting off a single collimator is shown by the dashed line in
Fig. 3. Cutting is accomplished using wire cut electric discharge machining. A clean
cut surface is obtained because the sheets are diffusion bonded. A collimator with
height L as depicted in Fig. 1 is thus completed (the view from the right and left
in Fig. 3 corresponds to (a) in Fig. 1). The height L of the collimator is determined
by the length when the device is cut as depicted in Fig. 3. An advantage of this manufacturing
method is that the height of the collimator can be machined to any value in the final
step. L is enlarged for applications requiring high wavelength resolution. Applications
requiring high speed can be accommodated using a lower L value.
[0040] A one-dimensional collimator was described above, and a two-dimensional collimator
according to the second embodiment of the present invention will next be described.
Only one hole 4 was opened in the metal sheet 1 in Fig. 2, but as depicted in Fig.
4(a), a multitude (six are depicted in the figure, but any number thereof may be included)
of rectangular holes 4 are formed in parallel at a prescribed interval in this embodiment.
The metal sheets 2 and pressing sheets 3 are sized corresponding to the metal sheets
1, and a collimator is manufactured by the same method as the one used to manufacture
the collimator depicted in Fig. 1.
[0041] This completes a two-dimensional collimator in which the cross-section corresponding
to (c) in Fig. 1 is converted to a cross-section such as the one depicted in Fig.
4(b).
[0042] In the collimator thus manufactured, the metal sheets 2 sometimes undergo heat deformation
and cannot be maintained parallel to each other when the holes 4 (vertical length
in the figure) depicted in Figs. 1 and 2 have considerable length. In such cases,
a grid is formed as a reinforcement to the holes 4.
[0043] Specifically, a linear portion with a width of approximately 100 µm is left remaining
in the hole 4 in the metal sheet 1 when a single hole 4 is formed by etching, as depicted
in Fig. 5(a), such that a narrow grid 6 is formed. When a collimator is formed by
such a method using this type of metal sheet, this collimator is formed such that
the cross-section corresponding to (c) in Fig. 1 is converted to a cross-section such
as the one depicted in Fig. 5(b). Specifically, a reliable collimator can be formed
without bending the metal sheets 2 because the grid 6 is formed so as to reinforce
the metal sheets 1 in this collimator. The light transmission efficiency declines
slightly because light does not pass through the grid 6 portion, but about three grids
are actually sufficient. Even when the width (vertical direction in Fig. 5) is 2200
µm and the grid 6 consists of three parts and has a width of 100 µm, the decline in
transmission efficiency amounts to about 3/22, which is not a significant problem.
It is apparent that the same method can be used for the aforementioned two-dimensional
collimator as well.
[0044] Because the portions of the metal sheets 2 not held on the either sides by the metal
sheets 1 (specifically, the portions corresponding to the holes 4) have low strength
(being the strength of a sheet about 40 µm thick) when wire cut electric discharge
machining is performed in any of the methods described above, these portions of the
metal sheets 2 sometimes deform as a result of heat deformation or stress during cutting.
The metal sheets 2 are shaped as depicted in Fig. 6 to prevent this phenomenon. Specifically,
holes 8 are formed in advance by etching or the like in the central portion of the
metal sheets 2 at positions aligned with cutting planes 7 formed by wire cut electric
discharge machining. The holes are sized so as to completely cover the width direction
(vertical direction in Fig. 6) of the holes 4 formed in the metal sheets 1 when the
metal sheets 1 and 2 are superposed to each other. Specifically, the height thereof
is equal to or greater than the width of the holes 4. In actuality, it is preferable
that the width of the holes 4 and the height of the holes 8 be substantially identical,
and that both types of holes be nearly completely superimposed over each other.
[0045] In this manner, the portion cut by wire cut electric discharge machining may extend
to both holes 4 and 8, and almost none of the portion of the metal sheets 2 forming
the side surface of the holes 4 undergoes wire cut electric discharge machining. This
portion thus experiences no heat deformation, and an accurate slit is formed therein.
[0046] In any of the collimators described above, the metal sheets 1 are connected at first
as a single sheet, but are divided into fine portions after undergoing wire cut electric
discharge machining. However, the divided portions in this step remain together because
they are bonded to the metal sheets 2 by diffusion bonding based on thermocompression
bonding.
[0047] Fig. 7 is a diagram depicting a structural outline of a spectrophotometer as an example
of an embodiment of the present invention. This spectrophotometer is a one-dimensional
spectral device, and is structured around a linear variable filter 11, a collimator
12, a linear sensor package 13, and a linear sensor 14. Light that is spectrally divided
by the linear variable filter 11 passes through the collimator 12 and is guided to
the linear sensor 14 inside the linear sensor package 13. The linear variable filter
11, collimator 12, and linear sensor package 13 are structured in contact with each
other. The linear sensor 14 is secured within the linear sensor package 13, and has
functionality for receiving light from the collimator 12 through a glass window (not
shown) and converting the light into electrical signals.
[0048] The wavelength of the light passing through the linear variable filter 11 is determined
according to the position in the width direction of the linear variable filter 11.
The collimator 12 guides the light emitted from each position in the width direction
of the linear variable filter 11 to the linear sensor 14 without mixing the light
with light emitted from other positions. The spectral characteristics of the light
entering the linear variable filter 11 can thus be determined by detecting the output
of each element of the linear sensor 14.
[0049] Emitted light from the linear variable filter 11 can be transmitted in a state of
low attenuation to the linear sensor 14, and sensitivity can be improved in the present
embodiment because the collimator 12 is configured to transmit emitted light from
the linear variable filter 11 to the linear sensor 14 with an air space as a medium.
[0050] Fig. 8 depicts the first example of a collimator used in the embodiment shown in
Fig. 7. In this example, the capillary plate described in the homepage of Hamamatsu
Photonics is used as the collimator. This capillary plate has holes with diameters
of a few micrometers to several hundred micrometers formed in a regular manner in
glass, and can be manufactured with a thickness of 0.5 mm to several tens of millimeters.
[0051] The capillary plate can be used as a collimator by applying a coating for complete
absorption of light to the inner surface of the holes in the plate. The spectral wavelength
emitted by the linear variable filter 11 is transmitted to the linear sensor 14 by
means of this collimator.
[0052] However, this embodiment has drawbacks whereby the aperture ratio of the capillary
plate is low, being 55% at the most, and the light transmission efficiency is also
low because the holes are circular. Particularly when the device used is a one-dimensional
spectrometer, the number of unusable portions increases because of the overall circular
shape of the capillary plate.
[0053] These problems can be overcome using a collimator such as the one depicted in Fig.
1, or a modified example (with a grid) of a collimator such as the one described above,
which are embodiments of the present invention. Specifically, because the openings
in these collimators are rectangular, the transmission surface area of light can be
enlarged for the capillary plate, thus making it possible to increase the light transmission
efficiency.
[0054] Fig. 7 depicts a one-dimensional dividing device, but if a wide device is used for
the linear variable filter 11, a two-dimensional spectral device can easily be configured
by converting the collimator 12 to a two-dimensional collimator, and constructing
the linear sensor 14 as a two-dimensional device.
[0055] In this case, even if a capillary plate is used for the collimator 12, drawbacks
are eliminated to a certain degree when compared with the case of a one-dimensional
spectral device, but a decline in light transmission efficiency cannot be avoided
inasmuch as the holes are circular. Consequently, the light transmission efficiency
can be enhanced when using a two-dimensional collimator such as the one depicted in
Fig. 4, or the modified example having the grid.
[0056] A one-dimensional collimator with a grid as depicted in Fig. 5(b) was manufactured.
SUS was used in the metal sheets 1, metal sheets 2, and pressing sheets 3; the metal
sheets 1 were 90 µm thick, the metal sheets 2 were 10 µm thick, and the pressing sheets
were 2 mm thick; and the holes 4 formed by photolithography and etching were 2200
µm wide. As depicted in Fig. 5(a), a five-member grid with a width of 100 µm was formed
at regular intervals within the 2200-µm space. As depicted in Fig. 6, holes 8 with
a width (horizontal direction in the figure) of 1 mm and a length (vertical direction
in the figure) of 2400 µm were formed in the center of the metal sheets 1 by photolithography
and etching at 4-mm intervals in the transverse direction of the figure. 128 metal
sheets 1 and 127 metal sheets 2 were alternately stacked on a pressing sheet 3, another
pressing sheet 3 was ultimately placed thereon, the assembly was diffusion-bonded
by thermocompression bonding according to the method described in the embodiments,
and the portion with the holes 8 opened therein was cut by wire cut electric discharge
machining. A collimator with a thickness of 3 mm, a width of 2200 µm, and a length
of approximately 15.8 mm was thereby completed. 128 holes 90 µm wide and 2200 µm long
are formed in this collimator.
[0057] The spectral characteristics of the one-dimensional spectral device depicted in Fig.
7 and obtained using this collimator were investigated using a device such as the
one depicted in Fig. 9. A parallel beam of light having a continuous spectrum emitted
from a parallel light source 21 is diffused by a diffusion plate 22 and transmitted
by a wavelength calibrating filter 23, and the wavelength distribution is measured
by a one-dimensional spectral device such as the one depicted in Fig. 7.
[0058] Fig. 10 shows the results of direct spectral decomposition of light emitted from
the parallel light source 21 without using the diffusion plate 22 or wavelength calibrating
filter 23; and Fig. 11 shows the results of spectral decomposition of light obtained
when a didynium filter is used as the wavelength calibrating filter without the use
of the diffusion plate 22. The theoretical light absorption peak wavelength for a
didynium filter is 580 nm, and the theoretical light absorbance thereof is 1.80, but,
as is apparent from Fig. 11, there is a peak at a wavelength corresponding to 580
nm (shown in the figure as pixel units of a linear sensor), and the light absorbance
thereof is 1.769. It is now apparent that the resolution of this spectral device is
particularly high. As previously mentioned, such high wavelength resolution is obtained
even when an interval of approximately 2 mm is provided between the collimator 12
and the linear sensor 14, because a collimator calibrated in the above-described manner
has extremely good collimating characteristics.
[0059] By contrast, the FOP and linear sensor had to be attached and the linear sensor package
13 had to be machined in the invention of "the prior application". In the present
embodiment, adequate characteristics are also obtained because of such characteristics
when the collimator 12 is connected to the linear sensor package 13.
[0060] A diffusion plate 22 was provided in front of the didynium filter, and the same type
of experiment was performed to further investigate the collimator performance. Fig.
12 shows the results of measuring the spectral characteristics obtained when the diffusion
plate 22 alone was provided, without any didynium filter. According to a comparison
between Figs. 10 and 12, both characteristics remain substantially unchanged regardless
of the predicted decline in wavelength resolution in Fig.12 because light enters the
linear variable filter 11 from various angles in Fig. 12. This is attributed to the
fact that because the collimator 12 has adequate collimating characteristics, only
rectilinearly propagating light is transmitted to the linear sensor 14 even when the
diffused light enters the linear variable filter 11.
[0061] Fig. 13 shows the results obtained when a didynium filter is introduced as a wavelength
calibrating filter 23 behind the diffusion plate 22, and light passing therethrough
is spectrally divided. An absorption peak also appears in Fig. 13 at a position corresponding
to 580 nm, which is the theoretical light absorption peak wavelength of a didynium
filter, and the absorption curve pattern is almost identical to the one depicted in
Fig. 11. However, the absorbance is 1.65, which is somewhat lower than in the case
in which the diffusion plate is absent.
[0062] As is apparent from the results above, high wavelength resolution is obtained using
the aforementioned collimator even when the light consists of diffused light, and
there is an interval between the collimator and the linear sensor.
[0063] The experiments described above are for visible light, but Fig. 14 shows the results
obtained by spectrally dividing transmitted light using an infrared light source as
the parallel light source 21, and using only a didynium filter without inserting a
diffusion plate 22. The absorption peak of the didynium filter for infrared light
is 800 nm, and the theoretical light absorbance thereof is 1.20, and there was an
absorption peak at 800 nm in the spectrometer as well, and the measured light absorbance
of 1.201 was extremely close to the theoretical value. It is thus apparent that the
spectrometer can also be used for the spectral decomposition of infrared light as
well.
[0064] Fig. 15 depicts spectral characteristics as a comparative example obtained using
a spectral device (depicted in Fig. 16) featuring the fiber optic plate described
in the invention of "the prior application", and using the diffusion plate 22 and
didynium filter. The light absorption peak for the dydinium filter absorption wavelength
of 580 nm declines to 1.30, and the waveform generally falls off. This is because
the collimator performance is poor in comparison with the embodiments of the present
invention.